Indian Police Expand AI Facial Recognition System, Arrest 70 Suspects in Delhi

Indian police forces are expanding their use of artificial intelligence-based facial recognition systems (FRS) to aid in criminal identification and tracking, following a nationwide trend of biometric surveillance that began with metro station deployments in major cities. The technology has been deployed across multiple regions, with Delhi Police reporting significant results in recent months.

Between September and November, the North Delhi police used FRS technology to arrest 70 individuals accused of theft and snatching. The system operates from a dedicated facility in Civil Lines, where a team of five officers manages the technology. A mobile unit equipped with four cameras captures images of individuals, automatically comparing them against a database of known offenders.

The FRS infrastructure can process footage from over 100 cameras simultaneously, performing live-camera face matches against one million records in 200 milliseconds. The system’s database incorporates images of known criminals along with footage from CCTV cameras and other sources, similar to the AI-based security systems deployed at national landmarks like the Red Fort.

“It has been acting as a force multiplier for law enforcement, providing a real-time edge in identifying people wanted for serious crimes, those subject to specific court conditions, habitual criminals, missing persons, women and the elderly,” said DCP (North) Raja Banthia.

Implementation results vary by district. Sarai Rohilla reported 16 solved cases using the technology, while Bara Hindu Rao documented 13 cases. Civil Lines reported four cases, and both Timarpur and Kotwali recorded three cases each.

The Digital Personal Data Protection Act (DPDPA) 2023 provides the regulatory framework for such systems, allowing government agencies to process data without consent under specific circumstances. However, the legislation has drawn attention for its broad exemptions and requirements regarding data accuracy. The Internet Freedom Foundation has called for a three-year moratorium on biometric facial recognition systems in India, citing privacy concerns.

Technical challenges include potential bias and inaccuracy in facial recognition systems, which can affect identification reliability. A recent data breach involving the Telangana Police highlighted additional concerns regarding database security and transparency in data handling practices. The incident, which compromised the TSCOP app, underscores the vulnerabilities inherent in large-scale biometric surveillance systems.

Sources: The Core IAS, Times of India, IDTechWire

—
December 25, 2024 – by the ID Tech Editorial Team